28.05.2024
On June 1-2, 2024, the OPEN DATATHON will be held at the Faculty of Economics and Business Administration, in room 401. The event will start at 9:30 AM and is open for student participation.
1) Participants create a solution for a selected task that applies some AI algorithm. Any programming languages and environments are acceptable (including graphical scripts like Scratch or a simple spreadsheet), as long as the program code (or blueprint) can be reviewed and can be run with ordinary efforts. The goal is to achieve a working product, even if it has limited functionalities.
2) The algorithms we recommend:
* Ant Colony Optimization
* Particle Swarm Optimization
* Reinforcement learning
* Generative Adversarial Network
* Neuro Evolution of Augmenting Topologies
* Genetic / Evolutionary Algorithm
* Multi-agent modeling
* Fourier decomposition
* Multi-stage Selection Procedure
* Artificial Neural Network
* Of course, it is acceptable to do something with other self-organizing algorithms/methods/simulations (for example, Boids simulation, Cellular Automata, morphs, or others).
3) The tasks described below are also only recommendations - everyone is free to propose another real or abstract task (using publicly available data), to be solved with this class of algorithms/methods.
* Edu Game on AI - Create an educational game, which is an example of AI itself. This task requires making a game that depicts an AI algorithm. (single-player, multi-player, zero-player) http://bit.ly/AI-games
* Drone detection through radio signals - Detection of drones through radio signals.
* NumerAI - Using some of the algorithms/methods, create a prediction system that produces a reasonably accurate prediction for any of the weekly Numer.Ai competitions. https://numer.ai/
* Quantiacs - Using some of the algorithms/methods, create a prediction system that produces a reasonably accurate prediction for any of the Quant competitions. https://quantiacs.com/ , https://www.quantconnect.com/
* Crypto Bot - Create a model for predicting the prices of major cryptocurrencies and an autonomous AI decision-maker for trading/investing.
* Portfolio Optimization - Create a dynamic portfolio optimization system.
* Real estate prediction - Create a model for predicting the real estate market. You may use the data from here: https://www.imoti.net/bg/sredni-ceni or any suitable data set (e.g., from Kaggle).
* Game Playing Bot - Create a bot that plays an already existing casual game. Open AI has published mega libraries of simple game environments (Gym and Universe), often used as testing grounds for AI algorithms.
https://blog.openai.com/universe/, https://github.com/openai/universe, https://gym.openai.com/, http://bit.ly/AI-games
* Covid-19 prediction - Create a model for predicting the spread of infectious diseases.
* Innoair Route - Create an optimization model for on-demand public transport - random virtual stops within a relatively small geographical area.
* Traveling salesman problem - Create an optimization model for the traveling salesman problem.
* Cooperative game bots - Create more than one bot (different bots could have different behaviors), which have to cooperate to beat a game. http://bit.ly/AI-games
* Personal data - Use a suitable algorithm to generate a synthetic data set that is statistically indistinguishable from the original data set (most likely personal data). You may use this data set: https://docs.google.com/spreadsheets/d/1wzlIVU62dBbQHLizhOatry6VYkdmNJ4H/edit?usp=sharing&ouid=116174241199284364326&rtpof=true&sd=true
* Artificial Life - Create an ecosystem of competing species.
https://www.youtube.com/watch?v=4iQUcGyQhdA
https://www.youtube.com/watch?v=r_It_X7v-1E
https://www.youtube.com/channel/UCKzJFdi57J53Vr_BkTfN3uQ
http://b3s23life.blogspot.com/2006_09_01_archive.html
http://www.red3d.com/cwr/steer/
https://github.com/davidrmiller/biosim4
* Kaggle - Use any of the AI algorithms to create a solution for any of the (past or present) Kaggle competitions. https://www.kaggle.com/
* Other - You are free to propose your own task.
4) Additional recommendations:
* Submit in the correct, working file format (e.g., .R instead of .txt or .py instead of .txt)
* The software should be self-contained - installation of all necessary libraries should be included
* List of library version requirements
* Independence from the operating system
* Documentation, or sufficiently detailed comments in the code
* Brief explanation of what is being solved and how (even in a comment)
* Acceptable clear formatting
* If possible: detailed visualization of the results
* If possible: define the entire working environment.